Traitement de liqueur Bayer synthétique par plasma submergé : étude d'impact sur la productivité

L'opération du procédé Bayer pour la production d'alumine introduit a chaque cycle une certaine quantité de molécules organiques. L'accumulation de ces contaminants est inévitable en raison de l'opération en boucle fermée dudit procédé. Une baisse de rendement et de productivité de I'usine est observée lorsque Ie niveau de contamination de la liqueur atteint un certain seuil. L'élimination de ces composes organiques devient donc nécessaire au bon fonctionnement de 1'usine. La recherche effectuée a couvert trois aspects distincts. Le traitement par plasma de différentes liqueurs synthétiques a été fait afin de déterminer la performance du précède en fonction des différents paramètres d'opération. Ensuite, I'effet de l'isophtalate et du phtalate en liqueur synthétique sur la productivité de celle-ci a été déterminé. Finalement, Ie traitement de deux liqueurs d'usines réelles a été fait dans Ie but de déterminer la performance du système en utilisant les deux liqueurs types sélectionnées. Le procédé novateur de traitement a été développé a I'Université de Sherbrooke en collaboration avec Alcan International. Ce procédé utilise une torche plasma a courant continu (DC) de 40kW en mode submergée pour effectuer l'oxydation des molécules organiques dans la liqueur Bayer usée. Le système permet Ie traitement de 15 litres de solution a une température de 150°C et a une pression de 4,4 atm (65 psi). De plus, l'opération du réacteur peut être faite en mode batch ou continu et avec différents types de gaz plasmagènes. Les échantillons recueillis sont analyses par potentiométrie, chromatographie ionique et certains par précipitation. La détermination du carbone organique total est également faite. Le lecteur notera d'ailleurs que les références au carbone organique total (COT) dans Ie texte seront reconnues sous 1'acronyme anglophone (TOC) dorénavant.Abstract: The Bayer process for the production of alumina, is a closed loop process that suffers from a buildup of organic molecules. These molecules contaminate the Bayer liquor resulting in lower plant efficiency. Removal of the contaminants is necessary to maintain plant operation. This research covers three distinct aspects of organic contaminant removal. In the first, plasma treatment of various synthetic Bayer liquors was performed.The process performance was determined as a function of operation parameters. Secondly, the decomposition of isophtalate and phtalate in a synthetic liquor was studied. Finally, system performance using real plant liquors was evaluated by using the plasma treatment method on two liquors imported from Alcan's plants. This innovative thermal plasma treatment process was developed at the Université de Sherbrooke in collaboration with Alcan International.The process utilizes a 40 kW, direct current (DC) submerged plasma torch to oxidize organic molecules in the contaminated Bayer liquor.The reactor may be operated in both batch or continuous modes with various types of plasma gases. Samples taken during testing were analyzed using the following techniques: potentiometry, ion chromatography, and some by precipitation.The determination of total organic carbon (TOC) was completed as well.The plasma system destroyed 22% of the isophtalate and 20% of the phtalate under identical conditions. For more concentrated Bayer solutions a loss in efficiency of approximately 10% is expected. Use of oxygen as a plasma gas increases the efficiency of the treatment. For the range of concentrations studied, it is not possible to clearly determine how each organic molecule interacts with other organic components in the liquor.The treatment of aliphatic molecules like adipate is more rapid than that of the aromatic molecules such as phalate. In fact, the difference between the kinetic constants from an oxygen plasma treatment is approximately one order of magnitude with k adipate = 0.0178/min and k phtalate = 0.0079/min. A study of the plasma gas volume as a function of the plasma feed gas established the importance of the solution recirculation rate on reactor oxidation efficiency. A higher destruction efficiency was also achieved during the first 3 minutes of the plasma treatment process. After the typical 30 minute test, the caustic concentration diminishes by approximately 20 g/l while the carbonate increases by 10 g/l.The productivity of the liquor varies from -0.10 g hydrate /gTOC phtalate and -0.27 g hydrate /gTOC isophtalate. This is on a rage of 0 to 18 g TOC /l of salt respectively. Degradation by-products generated during the plasma treatment are very undesirable and contribute to lower the amount of hydrate produced by the liquor. As an example, a liquor with 12 g/l TOC of phtalate, treated by the plasma process, will loose 8.7% of its hydrate productivity. A 12 g/l TOC liquor of isophalate will loose 8.9% of its hydrate productivity.The 0.2% difference results from different by-products being produced during decomposition. Plant liquors may be treated continuously using the submerged plasma. As expected, a residence time of 15 minutes is preferable to a 5 minutes residence time. At a 3 l/min flow rate, the treatment of a liquor with an initial TOC of 6 g/l is less efficient than a liquor with an initial TOC of 35 g/l.The productivity of plant liquors after plasma treatment varies. Each liquor has its own characteristics. For example the hydrate production of plant A liquor increases by 0.2 g hydrate/l for a treatment flow rate of 3 l/min. For the same treatment process, liquor from plant B looses 2.4 g hydrate/l in productivity.The seemingly contradictory behavior of the treated liquors may be a result of the initial composition. Plant B's liquor probably started with larger organic molecules. During treatment these were broken into a number of smaller organic molecules. Plant A most likely started with a lower molecular weight organics charge which were destroyed by the plasma treatment resulting in a reduction of the total low molecular weight organics.The production plant Bayer liquor tests have shown that the plasma treatment needs to be evaluated specifically for each Bayer plant before process implementation and scale up. Further work should also include study of the effect of other aromatic molecules on the liquor productivity. Tests with synthetic liquors/catalyst mixtures may also be useful for exploration of new reaction mechanisms. This may include using a reducing medium for organic decomposition. Further simplification of the synthetic liquor composition will permit a more detailed study and fundamental comprehension of this reactive system. Use of hydrogen peroxide as a catalyst is promising as well